I’ve spent the last 6 years of my working life obsessed with one question.
How can we harness tech to make the world a better place?
A few months ago, I delivered a lecture to 30 Masters’ students at University College London - my old university. The lecture became a synthesis of everything I’ve learnt about this question. A huge thanks to Alex and Duncan for the opportunity!
With minor edits, this piece puts those spoken words on paper (thanks to Jason for looking at a draft of this).
I was born in 1991, the year that marked the end of the Cold War.
Then, I spent the first 11 years of my life moving between Pakistan and the UK.
In my first year at university, the Arab Spring happened. Young people, fellow Muslims the same age as me, harnessed tech and social media to rise up against autocratic regimes in the Middle East and North Africa. Protestors used apps like Twitter, Facebook and YouTube to spread information about injustice, organise protests, and amplify their demands. I wrote my own Masters’ dissertation essay about it, right here at UCL in 2014.
I was optimistic. Tech, much of it from the Western world, could shape lives and communities for the better. It could reduce inequality between high-income countries, like the UK, and low income countries, like Pakistan.
Fast forward to the last few years, and we’ve had a global shift in how people think about tech. Social media makes us angry and hateful. Internet-enabled businesses like AirBnB and Uber erode communities1. Big tech avoids tax, invades our privacy, and manipulates our behaviour.
Most people now get that new tech ideas and products aren’t always a good thing.
So, the question then becomes: how exactly do we harness tech to make the world a better place?
When it comes to it, I believe that question nests three questions. They are:
The last six years at Brink have given me the privilege to act as grantmaker, coach, and strategist to amazing tech entrepreneurs and funders all over the world. Much of this piece draws from that well of experience, as well as my writing and thinking.
Let’s dive into each question, in turn.
I’ll cut to the punchline: Tech should give us the opportunity to shape our lives for the better.
That’s it.
If we want to make the world a better place through tech, that should be our ultimate goal. Not some other goal, which we assume on behalf of those we’re trying to help.
A few years ago, I worked with a Rwandese startup called Ampersand2 through the Frontier Technologies Hub3. Ampersand’s mission is to build and scale electric motorbike taxis in Rwanda, and then across sub-Saharan Africa.
The market opportunity is huge. There are 30,000 motorbike taxis in Kigali (the capital city of Rwanda) alone. Across East Africa, they cover enough distance to travel to the sun and back, twice a day.
Those bikes tend to be older models that are expensive to run and emit poisonous fumes into the air. Ampersand’s electric motorbike, by comparison, doesn’t emit any noxious fumes and charging them is much cheaper than traditional motorbikes and petrol4.
Think about what this technology unlocks for a taxi driver. They can work the same hours, dropping passengers from A to B, and make more profit. The downstream effects of this could mean the opportunity to pay school tuition for their children. Or building an extra room in their house for their parents. Or they could work fewer hours, and spend more time with their family and friends. It’s up to them.
And do it in an environment with cleaner air.
Ampersand’s electric motorbike taxis gesture towards the good all tech products should aspire to create.
If I were to sum that good up in one word, it would be agency. Tech should give us the chance to shape a life that means something to us, without assuming what that might be. It should give us agency.
This resonates with an approach to social justice and economic development developed by Amartya Sen and Martha Nussbaum, called the ‘capability’ approach5.
A ‘capability’ (as Sen & Nussbaum define it) is your freedom to live a life that goes well, as you would define it.
And crucially, this freedom has to be “substantive”. It’s not enough to be free from constraint. You should have the foundations to actually go out and do the things you want to do, and be the person you want to be6.
Any tech product claiming to do ‘good’ should provide the foundation to help you build the life you want (so long as that life doesn’t harm others), and then not interfere with your freedom to live it.
Sen and Nussbaum believe the primary goal of any policy or investment should be to grow capabilities. I’d extend this out to tech products too.
Any tech product claiming to do ‘good’ should provide the foundation to help you build the life you want (so long as that life doesn’t harm others), and then not interfere with your freedom to live it.
I’ve had the privilege to design and manage programmes and funds aimed at giving people agency. For instance, EdTech Hub7 was founded in response to the fact that 70% of children in the world, at age 10, can’t read and understand a simple sentence. It funds research and seeks to scale tech products to tackle this problem. The Assistive Tech Impact Fund8 is a grant-based scaling fund, founded because almost a billion people around the world need the use of assistive technology (such as eyeglasses, hearing aids, prosthetics, or wheelchairs), and 80% don’t have it.
Tech ideas that can provide an education, or ease a disability, undoubtedly give ‘capabilities’ to their users.
To scale those ideas to address unmet needs is to make the world a better place.
Which takes us to our next question...
Let’s start by unpacking a key word in that heading.
What do we mean by scale?
Here’s one visual that might pop into your head.
This ‘hockey stick’, popularised by Silicon Valley’s tech startup scene, is the globally dominant narrative around scale. A company tinkers (step 1), iterates, and develops its product (step 2). Then they find product-market fit. At this point, the startup ‘takes off’ (step 3) and enjoys ‘surging growth’ (step 4).
If the horizontal x-axis on that visual is time, the y-axis could be # of users, or $revenue, or gross or net $income.
This dynamic creates the unicorn: a company with a billion dollar valuation, because of the promise of surging growth across all these y-axis metrics.
Last year, I had the privilege of designing and delivering the scaling strategy curriculum for the 100x Impact Accelerator, run by the Marshall Institute at London School of Economics (LSE)9. 100x Impact’s mission is to nurture the next generation of social unicorns. These are startups that can tackle the world’s most pressing challenges in health, climate, and education on the scale of billions.
When it comes to having impact at the scale of billions, social unicorns needn't take the path popularised by Silicon Valley. They don’t have to grow their own company’s number of users, revenue, or income. They can have unicorn-level impact, without the huge company size or valuation.
And they can do that, by picking the right endgame.
I first came across Gugelev & Stern's 'endgames' concept on the 100x Impact Accelerator. Endgames are different routes to scale and impact for startups. Crucially, only one of these is “sustained service” - where the startup grows in size and directly provides a product and service to billions. Other endgames include:
I spent a lot of time talking about endgames with startups on the 100x Impact Accelerator. The big takeaway was: to build agency on the scale of billions, you don’t have to scale your own organisation. You can build a product, and others can deliver it or replicate it in their markets (like M-PESA Africa or kindergartens). You can certify, lend expertise, train or partner with others (like Marie Stopes International). You can have an idea, or some evidence, and reach people through advocacy, storytelling and influencing policy (like the Harvard School for Public Health).
Of course, it’s not as easy as it sounds. There’s a lot of grit and experimentation that comes with scaling an idea towards an endgame.
Here are three stories on scaling (involving plenty of grit) from organisations I’ve had the privilege of working with.
Scaling story 1: Bicycles in Zambia
Zambia is three times larger than the United Kingdom, but with three times less people. It’s large and people are spread out. What’s more, as of 2017, only 17% of the rural population lived within 2km of a good road13. Getting goods or people from A to B is difficult.
I had the joy of working with an awesome duo in Zambia: Wyson, the Founder of Onyx Connect (a local startup) and Steve, a civil servant for the UK Foreign, Commonwealth & Development Office. Steve had the idea of ‘pay as you go’ bicycles to help rural farmers in Zambia make long journeys. Wyson brought it to life, providing rural farmers with a bike, solar lamp, feature phone and GPS tracking unit for $5-10 per month (over 12-36 months).
Onyx Connect sold the first hundred bikes themselves. But their biggest unlock came when they partnered with companies and farming cooperatives. Partners marketed the product, or brought packages in bulk, for their employees or members. They also helped ensure timely repayment.
To date, Onyx Connect has sold 3,000 bikes. Their work has been featured on the BBC and Guardian. There’s more demand for pay-as-you-go bikes than they can keep up with, so they’re looking for loans or grant funding to scale their operation.
What’s even more promising is that other bicycle manufacturers are now investigating replication of the ‘pay as you go’ bike model in other countries.
On their own, Onyx might reach thousands. But by leaning into commercial adoption and replication from others, they might scale to millions of people.
Scaling story 2: Prosthetic arms in Sierra Leone
From the first moment I met Nate and Niall, I saw their obsession with building prosthetics like trainers. Koalaa - their startup - had recently won a Tommy Hilfigher fashion award for the design of their prosthetic arm. Why, they asked, couldn’t prosthetics be as trendy, comfortable and well-liked as a trainer (sneaker)?
Not only did the Assistive Tech Impact Fund buy into this vision, but they also saw Koalaa’s potential for amputees in sub-Saharan Africa. At a cost to Koalaa of $75 per year (including ongoing care)14, prosthetic arms could be transformative for users, at low cost. The biggest game-changer was that anyone could be certified to fit someone with the prosthetic after one day of training.
I worked closely with Nate, Niall and the Koalaa team for six months. We focused on Sierra Leone, a country where civil war left a legacy of persons without limbs.
On a whiteboard, we visualised the $75 cost to Koalaa per prosthetic as a pie chart. How could we meet it? In a memorandum of understanding with Sierra Leone’s government, they pledged to cover ~$20 of the cost through tax relief and access to their network of clinicians to fit the prosthetic. Wealthier amputees could sometimes cover all or part of the cost. UNICEF, the UK Foreign, Commonwealth & Development Office and other donors, charities or philanthropists might cover some of it too. And, Koalaa looked into whether sales in high-income countries might cross-subsidise sales in sub-Saharan Africa.
We worked backwards from an endgame. Koalaa retained control over the product’s award-winning design. The distribution, fitting, and ongoing care would be handed over to the government’s clinicians, or other certified bodies. A blended funding model of government support, philanthropists, donors, charities, customers and Koalaa’s own global profit would sustain the service.
I can’t think of a single product that enhances agency as much as a prosthetic. With it, you can eat, work, build, type, and hold. Koalaa is well on its way to validating a model and endgame that could bring such agency to persons with disabilities across an entire continent.
Scaling story 3: EdTech Zanzibar
EdTech Hub is a global partnership that supports donors, philanthropists and governments to make evidence-led investments. First, the Hub designs and conducts research. Then, it tries to shape the millions of dollars that gets invested in education technologies in low and middle income countries (LMICs). Like the smoking ban I mentioned earlier, it’s about using evidence, advocacy and storytelling to shape policy and big investments. Our ultimate goal is to make sure edtech spending helps teachers to teach, and children to learn.
I worked to shape investment in virtual learning environments (VLEs) by the World Bank and government of Zanzibar15. When rolled out, all secondary school children and teachers would use the VLE to access learning and teaching materials.
Our hope: it might make a dent on Zanzibar’s 40% secondary school dropout rate16.
We started by building an evidence base on how a VLE might work in Zanzibar. This included a feasibility study and guide to VLE content curation. Using this evidence, we built a “paper prototype” - an A3 sheet showing a blueprint for how the VLE would function in practice providing education resources to teenagers. We took this to 1:1 interviews with 24 teachers. 90% of those teachers told us their biggest problem was actually the inadequacy of educational resources available to them (textbooks, lesson plans, etc).
"No-one has the step-by-step blueprint on how to scale ideas like that, because no-one has. That’s why we work in experiments."
We went back to our paper prototype, iterating it to focus on content that teachers could use as teaching aides with their students17. This content would be available offline and include guidance for teachers on when to use it. The World Bank and government took this basic blueprint, and embedded it in their six year, $50M education spending plan.
Our idea, in this case, was an iterated paper prototype demonstrating how VLEs might work well, that scaled via government policy. It will - we hope - support thousands of Zanzibar’s teachers, and millions of children, giving them the means of shaping their lives.
Pay-as-you go bikes in Zambia. Prosthetics in Sierra Leone. Virtual learning in Zanzibar. Like I said, scaling ideas like this takes grit.
Wyson, Steve, Nate, Niall, and others were all working through things that hadn’t been done before. No-one has the step-by-step blueprint on how to scale ideas like that, because no-one has.
That’s why we work in experiments.
Across our portfolio of funds and ideas, from day one we ask: Can this idea have impact on the scale of billions?
Since it’s impossible to answer definitively, we gather data that signals whether our idea will scale. And we do that as rigorously and as often as we can before running out of time or money.
Over the last 10 years, ‘lean’ approaches have made their way into the world of startup, and then into the world of scaling ideas that tackle entrenched social problems.
‘Lean’ as a business term started in the manufacturing sector in Japan. Companies like Toyota adopted lean manufacturing to mean two things. First, identify what brings value to a customer. Then, deliver this value, with no waste, and continuous improvement of products.
Eric Reis adopted these same principles for the world of startups, and later, large corporates. For Reis, startups should build the smallest possible version of their idea, test it with customers, learn, and iterate. Getting through these cycles quickly, for Reis, the only good use of a startup’s limited resources.
For an idea seeking social impact at scale, I believe it’s the same story.
At Brink, we refer to these cycles of building, testing and learning as experiments. By now, I’ve run hundreds of these experiments. They always start with the same three questions:
Here’s two experiments with two of the ventures I’ve mentioned already, Ampersand18 and Koalaa19.
Eric Ries talked about two types of uncertainty a startup faces. Uncertainty about value, or whether a product delivers enough benefits to a customer so they’ll pay you (that’s what Ampersand’s experiment was testing). And uncertainty about growth, or whether a product can reach more users (that’s what Koalaa’s experiment above was testing).
Ann Mei Chang, who has adapted Eric Reis’ ‘lean startup’ for social impact, adds one other type of uncertainty. That’s uncertainty about impact, or whether a product actually improves lives. Or, as I would put it, uncertainty about whether the product helps people to shape the life they want.
One more nuance: For many startups we work with, the person who uses the product is different to the person who pays for it. Take Koalaa, an amputee uses the prosthetic, while different actors (government, donors, charities, etc) pay for it. So, uncertainty about value needs to cover value to an end user, and value to different payers.
The first step is to express exactly what you are uncertain about as a belief. The second step is to test it. The third step is to iterate your product and approach based on what you learn. Systematically working through this process is the hard work of scaling ideas.
Let’s bring this together with one final example. During my time on the Frontier Tech Hub, Antony H. - a civil servant and true intrapreneur - approached us with the idea of using ‘smart’ (internet connected) vending machines to distribute contraception in Kenya.
His impact assumption: 47% of pregnancies are unintended among girls aged 15-19. What a difference discrete, timely access to contraception (through a well-stocked vending machine) could make on these girls’ and their ability to live the lives they want20.
We21 convened our target user and asked them whether they would use a vending machine. Later, we built a prototype and invited them to use it. They were eager to, as long as two things were true. One, the machine hooked up to M-Pesa (Kenya’s widespread mobile payments system, mentioned earlier). And two, they spent as little time as possible with the machine (lest someone catch them using it). This validated our value assumption to the end user.
But what about the value to donors, who would have to fund the machines in very low income, remote parts of Kenya?
Fortunately, both the Foreign, Commonwealth and Development Office and the THE CHILDREN'S INVESTMENT FUND FOUNDATION (UK) (CIFF) had family planning as a key priority. After several meetings, we unlocked £150K of co-investment from CIFF alongside the initial £100K from the Frontier Tech Hub. Both organisations also committed time from a ‘point person’ to follow progress. This signalled that they felt the idea could deliver value to them.
A critical belief around growth was whether Kenya’s Ministry of Health and Pharmacy & Poisons Board would allow us to deploy the machines in public places. Spoiler: They did not.
We were permitted to place vending machines in licensed pharmacies or privately owned places only. The government spoke for a socially conservative country, where many disapprove of contraception. To scale to more users, we would have to grow slowly. This belief - that government agencies would allow us to place vending machines in public places - became our most critical.
More on how this story ends, later.
What I want to show with this story for now, is how beliefs about value, growth and impact connect, and the hustle of testing and learning across them, as often and quickly as you can.
So far, we’ve talked about testing and learning at the level of an idea. Because you don’t know in advance what will scale (and how), you need to spend time and money on experiments, to learn what works before spending more.
This same principle applies if you are a donor, philanthropist or investor.
Venture capitalists (VCs) seeking a financial return on investment know this. That’s why they typically reserve 40-60% of a Fund, to invest in future funding rounds for their portfolio companies. It's also why they invest across many companies22 in the knowledge that most won’t scale, and the hope that those that do will ‘return the fund’.
Impact-focused funding is following suit, where it funds high uncertainty ventures. For instance, the US Agency for International Development’s Development Innovation Ventures’ (DIV) has a tiered approach to funding. It provides $200K, then $1.5M, and then up to $15M of grant funding. As ideas validate more beliefs, they access more funding.
Catalytic philanthropy, popularised by Bill Gates among others, promotes the concept that grant capital can “get things started” by de-risking uncertain early-stage ideas for governments or markets to adopt and fund at scale23. Catalytic philanthropists, like VCs, need to take many swings before the home run. DIV, for instance, has funded 297 separate ideas in 52 countries between 2010-24.
If anything, catalytic grantmakers need to take even more swings than traditional VCs. Not only are their portfolio companies validating beliefs around value and growth, but impact too.
Over the last year, I would say I’ve worked hands-on with 40+ ventures. Four that I’ve mentioned - Ampersand, the VLE in Zanzibar, Koalaa, and the pay-as-you-go bikes in Zambia - have, by my estimate, improved lives more than the bottom half of the portfolio.
I come back to core concepts like agency, scaling, endgames, and experimentation a lot. But the school of thought I’ve come back to most in the last five years is Organisation & Relationship Systems Coaching (ORSC).
ORSC provides an approach and toolkit for managing the dynamics in teams. It's core insight: every team is a ‘third entity’, with a personality, values and wisdom of its own.
"When it comes to scaling tech ideas, the hard work isn’t in selecting an endgame or designing an experiment. It is in understanding the team and nurturing its resilience, creativity and psychological safety in order to make progress. "
When it comes to scaling tech ideas, the hard work isn’t in selecting an endgame or designing an experiment. It is in understanding the team (the ‘third entity’) and nurturing its resilience, creativity and psychological safety in order to make progress.
Building on ideas such as ORSC, we’ve developed behavioural innovation at Brink. Behavioural innovation is our proven method, combining innovation methods with insights from the behavioural sciences on incentives, psychology, and change.
Take experimentation. If we’re running good experiments, at some point we’ll face data that compels us to do things differently. When we do, we’ll likely have to navigate a host of behavioural barriers. Things like:
Every team will have its own, unique interpersonal dynamics. These might make it hard for some to speak up, or make conflicting ideas more or less likely. Whatever it is, these factors - sometimes erroneously called the “soft stuff” - are the meaty, hard work of scale and impact.
We’ve dived into the nuts and bolts of scaling game-changing ideas.
But how can you make sure the idea you’re scaling actually makes things better in the long run?
Like I said, social media seemed to promote human flourishing and connection. It was after it scaled to billions of users that most people started to feel it has a net-negative impact. Rather than making us more free to shape the lives we want, it's seen as a tool for oppression through fake news, disinformation and doomscrolling.
While you can never be totally sure what the answer is, this last question is way too important to leave to chance.
So here are three rules of thumb I’ve built up over the last six years. Think of them as a shorthand, improving the odds (but not guaranteeing) that a tech idea will really enhance agency and make the world a better place.
1. Build from the bottom up.
Earlier, I mentioned our pay-as-you-go bicycles in rural Zambia. Farmers pay in small increments, taking full ownership of a bike, solar lamp, feature phone and tracking GPS unit over 12-36 months.
Our initial idea was very different. Inspired by schemes in Europe, we thought users might ‘rent’ a journey, paying a small amount per hour to ride a bike from A to B.
It was Wyson, Zambia-based Founder of Onyx Connect, who taught us that renting anything was an alien concept in rural Zambia. If you pay for something, you should, one day, own it. I’ll never forget when one bike shop owner told me how his customers bought bike parts over time - first the frame, then the wheels, then the handlebars (and so on). They did that rather than renting a whole bike. Ownership mattered, much more than in urban Europe. This insight pivoted our entire model, towards one where users could pay in increments and eventually own the bike.
This new model worked with the meaning people in a country built up for themselves over time, rather than importing a mental model from somewhere else. The idea was built from the bottom up.
Again, this comes down to the concept of agency. If people value something, then doing ‘good’ is to enable them to live a life by that value.
And working with locally based Founders like Wyson, rooted in their context and community, is the single best way to do that. They’ll understand the context: the behavioural, psychological, emotional, cultural and linguistic dynamic into which a product lands.
It’s not binary. We can import tech and new ideas into new contexts, while aligning them with local needs and values.
Between 2019-20, Frontier Tech Hub and UNICEF collaborated to fund Swoop Aero, an Australian drone startup, in Malawi. Our goal was to test whether multi-purpose drone operations were cost-effective and viable in rural Malawi.
Introducing drones into a new country can be disruptive. People on drone flight paths will experience a noisy, alien object. Drone deliveries might also take away income from local couriers. Images from drones, taken to support agriculture or monitor infrastructure, might invade privacy.
Marie-Claude, our partner from UNICEF, and Martyn, who led African operations for Swoop Aero, recognised these disruptions. They responded by holding our work accountable to a key metric: the number of Malawians trained as drone pilots or operators24. Swoop Aero would provide the cutting edge tech; Malawians would shape its arc in their own country.
"Introducing our idea slowly and building acceptance over time would give our idea a much better shot at becoming part of the fabric of a place, and gradually, sustainably scaling within it."
Or take the story I started earlier, on smart contraceptive vending machines in Kenya. I mentioned that Kenya was a socially conservative country where a majority don’t believe in contraception.
Our locally-based partners - Population Services Kenya and Novek - recognised this and recognised that non-interference was not an option. As I said, 47%(!) of pregnancies among adolescent women are unintended. And unintended pregnancies can be lethal: over 1 in 200 births lead to a woman’s death in Kenya25.
To square this circle, we introduced a new term: scaling gently. We faced an acute, urgent crisis but knew that muscling in and scaling a tech solution would rub against deeply held values. Introducing our idea slowly and building acceptance over time would give our idea a much better shot at becoming part of the fabric of a place, and gradually, sustainably scaling within it.
For instance, we started in student halls and campuses. Although universities weren’t where the crisis was most acute (most students used contraception already), they were sites where the stigma against contraception was less prominent.
Starting here gave us the chance to learn more about how the machines would remain stocked, understand user behaviour, and iron out technical issues.
We also stocked the machines with crisps and drinks, as well as condoms and morning after pills. Again, this was part of scaling gently. By selling things people would have seen in other vending machines, ours stood out less.
Over time, we built up trust in the contraceptive vending machine. Eventually, Kenya’s Ministry of Health and Pharmacy & Poisons Board granted us permission to operate in public spaces. Even then, we started with bars and restaurants before moving onto bus shelters and public roads, where we knew our machines would be most controversial.
I hope these examples make one thing clear. By introducing and scaling tech, you are introducing and scaling a set of values embedded in the tech. Tech is never, as people sometimes say, a ‘neutral’ tool. That’s why keeping agency to those rooted in the places you work is an ethical imperative; it minimises the gap between values embedded in the tech, and values of the community receiving the tech.
Before moving on, I want to address one huge barrier to locally-led tech: the distribution of different types of capital.
Venture capital (VC) funding fuels early-stage tech. It’s the risk-acceptant funding entrepreneurs need to experiment with game-changing ideas. But the gulf in availability of VC funding available in different parts of the world is stark. Last year, North America saw 11,480 VC deals totalling $114.3bn. Europe saw 6,905 deals totalling $52.4bn. Africa, on the other hand, saw 545 deals totalling $3.6bn. Latin America saw 489 deals totalling $3.0bn26.
There’s less funding available, and less per deal for new ideas in emerging markets like Africa and Latin America. Which means less time to iterate towards the scalable version of your idea. Talented entrepreneurs outside of the US/Europe, and those whose lives would be improved by their products, are systematically neglected by VC.
In one universe, overseas aid might counterbalance this. Total international development spend in 2023 was $223.7bn27 - almost 2x VC spend in the US.
However, unlike VC, most of this funding is earmarked for things with fixed, predictable outcomes. Donors find it difficult to justify using even a small portion of this money for testing new ideas and technologies. Almost all of it ends up with large NGOs and implementers who build roads, schools, and clinics and execute tried-and-tested interventions. These interventions, while important, don't have the high upside potential of new, unproven ideas.
And almost all of it ends up channelled through organisations in the Global North.
Over the last 6 years, I’ve seen these dynamics start to shift.
Catalytic philanthropy, coined by Bill Gates (among others), reimagines global aid as a space to place bets that governments or markets might see as too risky.
Localisation, a movement started in humanitarian response, has broadened to advocate for more funding to organisations based in the same countries as end users.
Impact investment, pioneered by organisations like Acumen, is applying the VC playbook of taking equity ownership in high-uncertainty ideas that might generate outsized financial and social return.
Diving deep into these trends would be another lecture. For now, let’s hope they continue, and fuel a generation of new, game-changing, locally-led tech ideas.
2. Consider the unintended consequences
Above, I presented the case for tech imposing values in a place.
Here, I want to highlight another imposition: new tech ideas impose new problems.
Cars have given us mobility, but the frustration of traffic jams and upwards of 1.3 million deaths a year. Healthcare advances raise our life expectancy, but leave us facing more of the aches and pains of old age28. Microsoft Teams and Slack solve for collaboration across different locations, but fill our lives with noise, and always on work talk.
Edward Tenner, in Why Things Bite Back29, refers to these new problems as technology's ‘revenge effects’. In Tenner’s words:
"Technology demands more, not less human work to function. And it introduces more subtle and insidious problems to replace acute ones."
By introducing “subtle and insidious” problems into peoples’ lives, tech reduces their agency. Such problems compromise your ability to pursue the good life, as you would define it.
I saw revenge effects play out in Zimbabwe, with one of the Frontier Tech Hub’s very first early-stage grantees: Africa Power Storage (APS)30. APS saw the potential of off-grid solar energy systems to provide reliable energy to health facilities like clinics and hospitals. But too often, solar power gets installed, then forgotten about. A fault develops, or the panels provide less and less power, and stop working at all.
With 3G/4G connectivity increasingly prevalent (including in rural areas), APS saw the potential of connecting solar energy systems to the internet. In other words, making them 'smart' like the vending machines we saw earlier. A smart solar energy system can send back data on what needs fixing, and data on usage so APS could ‘right size’ future solar energy systems.
The tech product here is complex: solar panels, an inverter, and a SIM card to connect to 3G/4G. And it solves an acute problem for health clinics. They can remain open after sunset, and emergency procedures such as childbirth can take place with electric lighting (not candlelight).
At the same time, I saw firsthand the new jobs this smart solar energy system created. When I visited a health clinic in Jari, western Zimbabwe, I saw that staff at the clinic now had to monitor energy capacity to ensure sufficient power, run basic maintenance, and keep the panels free of dust. The benefit of electricity was immediate; the new tasks it introduced took some time to reveal themselves.
In this case, those tasks were worth the benefits. The impact of energy on people’s lives was unambiguously positive on balance.
But this work did highlight that introducing technology is above all a responsibility. It might solve problems, but it also imposes new ones. The responsibility is to think through these revenge effects, particularly where they're unintended or non-obvious.
The responsibility is also to experiment with ways of minimising the knock-on effects, and to make the trade-off clear to your users.
Tech is not - and never will be - a ‘silver bullet’ that can solve problems cleanly and completely.
3. Remember that tech exacerbates inequality
Here’s a pattern I’ve seen over the last six years: technology will benefit you more, if you’re already better off. You could be better off in terms of personal wealth: your education, income, capital, and network. Or, better off in terms of the wealth of the system around you: your household, community, and country. Either way, you're better placed to make the most of tech products.
Last year, EdTech Hub partnered with EIDU, a personalised learning app and startup providing literacy and numeracy exercises to 300,000+ pre-primary students in Kenya.
The EIDU team31 has cracked two of the most difficult puzzles in scaling an edtech product in sub-Saharan Africa. First, they've built something that complements (not replaces) the teacher in a classroom. And second, they've placed a product within a government’s education system.
Our goal was to gather evidence on how EIDU have implemented digital personalised learning in schools and classrooms, and scale this evidence to governments, funders and edtech startups around the world.
As we were doing that, we found something interesting. In schools, higher performing children used the EIDU app much more frequently. And a good ~20% of children didn’t use it at all in a given week. Sometimes, this was because higher performing children were more curious about the app. Other times, the teacher let them use it so she could focus her energy on lower performing children. For some, their parents or older siblings dropped them off early or picked them up late, so they could use the app.
Whatever the reason, this opened up a dangerous prospect: EIDU might widen the gap between high- and low-performing children. The impact might be net-positive, but concentrated on those children who didn’t need it as much.
Working alongside Indrah from EIDU’s Kenya office, and the teachers who used the app every day, we prototyped product features that might lead to a more even usage. This included:
We found that when these features were introduced together into 20 schools over one term, equality of usage improved by 16%32. It has a particularly strong effect in classrooms where equality of usage was lowest.
Ensuring tech benefits all requires deliberate design. And this matters, because technological progress, left to its own devices, has the least impact on those whose agency is most at risk.
Our work with EIDU highlighted one way in which tech benefits those with a pre-existing advantage. But we see this in our own lives too. If you live in London, on a reasonable income, you probably have a newish smartphone and access to the latest version of all apps, with the 4G/5G and digital literacy to use them whenever and wherever you want. Now, compare that to someone in rural sub-Saharan Africa.
Who benefits more from technological progress? Whose ability to shape a life that means something to them is enhanced more by tech?
I started this lecture with three questions that anchor the notion of tech for good. Over the last 45 minutes, I’ve tried to gesture at answers. Here’s the one slide summary.
But if there’s one thing I hope you take away, it’s not what’s on this slide.
It’s this: We’re all alive to shape the course of technology ideas, and shape them towards good. When we do that, there’s perhaps no single bigger engine for positive change.
But it won’t happen by itself. It requires grit, rigour and responsibility in equal parts.
Onwards!
1. To elaborate: AirBnB turns long-term housing into short-term rentals, driving up rents and changing the face of whole neighbourhoods (so much so, that New York City has ‘de-facto banned’ it). Uber takes money out of public transit, increases traffic congestion, and routes money out of local communities. AirBnB, and Uber is banned in.
2. A huge shout out to Josh and Alp, Ampersand’s co-Founders. At the end of last year, they raised a $19.5m Series B round.
3. The Frontier Tech Hub is a programme funded by the UK Foreign, Commonwealth & Development Office and delivered by Results for Development, DT Global and Brink. It’s been running for 8 years, and provided funding and support to 70+ ventures.
4. As part of their offer, Ampersand provides both the electric motorbike and access to battery swap stations. Drivers pay in instalments, over several years.
5. The Stanford Encyclopedia of Philosophy has a great summary of the ‘capability approach’.
6. Many others have had similar ideas. Isaiah Berlin, for example, developed two concepts of liberty. “Freedom from” constraints on its own isn’t enough: we must also have the “freedom to” act on our own free will and realise our potential.
7. The EdTech Hub was founded in 2020, and funded by the UK Foreign, Commonwealth & Development Office, Gates Foundation, UNICEF. It is delivered by Results for Development, Jigsaw, University of Cambridge, Open Development & Education and Brink.
8. The Assistive Tech Impact Fund was active from 2020-22. It was funded by the UK Foreign, Commonwealth & Development Office, and delivered by the Global Disability Innovation (GDI) Hub, BFA Global and Brink.
9. A huge shout out to Nell, Leslie, Nickil, and the current 100x Impact team, who keep taking that accelerator from strength to strength 💪
10. Read more about Marie Stopes International’s social franchising model in this report.
11. Data from Khan Academy’s annual report (2022-23).
12. Read more about the fascinating history of the ‘designated driver’ campaign in this case study produced by the Harvard School of Public Health.
13. Data from World Bank Group.
14. This is far cheaper than the status quo. Typically, even basic prosthetics cost thousands of $$ to manufacture, fit, and aftercare.
15. None of this work would have been possible without Caspar, Tom, and Daniel on the EdTech Hub, and Raya from Cube Zanzibar Cube (Zanzibar’s first innovation hub) who collaborated with us for the project. Director Khalid, Abdalla, Masoud, Khamis, Hussein, and Mfaume were champions of this work within the Government of Zanzibar, alongside Tanya and Kaboko from the World Bank.
16. From EdTech Hub’s virtual learning environment feasibility study in Zanzibar (2020).
17. This cycle of getting evidence, applying it to an idea, testing the idea, and iterating it is at the core of EdTech Hub’s Sandbox method, which I developed alongside Lea, Alice and many others at EdTech Hub.
18. A note on Ampersand’s experiment: this type of experiment is called a “fake door”. It was popularised by Dropbox, who made a 4-min explainer video and asked users to join a waitlist, before they had developed their product. The waitlist was a “fake door” because there was no product behind it (yet), but it validated demand. Here’s the original video.
19. A note on Koalaa’s experiment: A lot of experiments I run are inspired by The Mom Test by Rob Fitzpatrick. In a nutshell, it encourages startups to be wary of compliments and seek meaningful commitment from users in the form of time, money or reputation. In this example, Koalaa were seeking commitment in the form of referrals, rather than meaningless assurances of working together.
20. Data from Adolescents’ narratives of coping with unintended pregnancy in Nairobi’s informal settlements (Numah et al, 2015).
21. Zahid from Novek (a local internet of things startup), and Wycliffe & Christine from Population Services Kenya (one of Kenya’s largest health NGOs) came together to build and test the idea.
22. Data from Follow On in Venture Capital (M. Palank & A. Srivastava, Going VC, 2021).
23. The power of catalytic philanthropy (B. Gates, 2012)
24. This metric would have remained an aspiration only, without UNICEF’s pioneering Drone and Data Academy in Malawi.
25. Data from the World Bank Group.
26. Data from the Africa Venture Capital Association (AVCA).
27. Data from the Organisation for Economic Development & Cooperation (OECD).
28. See also: more innovative ‘super shoes’ for running also increase injury risk, according to one study from 2023.
29. Published by Tenner in 1997, but more relevant than ever today.
30. Africa Power Storage was founded by Mike Rosenberg, a true tinkerer and innovator I had the pleasure of collaborating with. I wrote about my time in Zimbabwe, alongside Mike and his partners here.
31. EIDU has a formidable team based in both Germany and Kenya. Through my role as EdTech Hub’s Experimentation Lead, I had the pleasure of working with Bernd, Nina and Aidan in Berlin, and Indrah, Allan, Tess, Lilian, and Silas in Nairobi and Mombasa.
32. Further details on this study, including how we used standard deviation to measure equality of usage, will be published in March 2025.
